FIRST Extra-Galactic Surveys

Practical Considerations

Seb Oliver

Choices in Designing a Survey

• Bands

• Depth

• Area

• Fields

Depth

Depth

• Source confusion places strong constraints on depth

• Classical confusion limit from Condon 1974

• Super-resolution can improve things but is very expensive

Classical Confusion (Condon 1974)

1

1

effective

1

13

3

k

q

effective1

2 1

31

q

nq

df1

effective )],([

effective1

23

1 q

nq

kSdS

dnPower-law counts 2/5

Euclidean

D = q

D

dnS0

2

Classical Confusion

21effective 4

1

2ln)1(

1

2

effective 2.118.0

D

2

effective 2.11

27.0~

D

Effective beam for Gaussian Profile

Effective beam for Airy Profile

rough approximation assuming same functional form as for Gaussian

if= 2.5

Example: ISO HDF South

25 arcmin 2.2 n

1'5 circle radius5.243

N

23effective arcmin106)μm15(

c.f. ~30 Oliver et al. 2000

UK-SCUBA 850 m Surveys

Hughes et al. 1998 Nature 394 241

Classical 5confusion limit 0.43 sources arc min-2

Area = 8.7 arc min2

5limit 3.8 sources

c.f. 5 sources in

Classical Confusion Limits from FIRST

m 70 120 175 250 350 500D/m 3.5 3.5 3.5 3.5 3.5 3.5 /arc^2 13.9 40.7 86.6 176.8 346.4 707.0n5 12469 4243 1995 978 499 2444.3 0.74 3.2 11 18.6 20 16.6

Last row is flux at which number of sources hits the 4.3confusion limit threshold using models of Rowan-Robinson 2000 ApJ in press

Limits to Super-Resolution

• Point Sources vs. Extended Sources

• 2 Point Sources vs. Extended Source

4/1NSuper

Natural

8/1NSuper

Natural

Lucy 1991Proc. 3rd ESO/ST-ECF data analysis workshop eds Grosbøl & Warmels

Lucy 1992 Astron Astro 261, 706

Lucy 1992AJ, 104, 1260

Perfect Instrument: records exact position of Ideal reconstruction

Limits to Super-Resolution (ctd.)

• Assuming super-resolved profile can be considered the same shape

16/18/116/18/1effective

ttNN

)1(16/1)1(8/1)1/(1effective

tt

24/112/1 tt

16/18/1 ttn

e.g. moving from a 15 minexposure to 100 hours wouldincrease the number of sources at the idealised super resolved confusion limit by between 2.1 - 1.5

reducing the confusion noise by 1.6-1.3

effective1qn

Area

Choice of Area

• Major survey projects are going to be large area classical confusion limited surveys– Large to detect rare/high luminosity objects– Large to produce statistically significant sub-

samples– Say 100 square degrees

• Niche projects over smaller areas may attempt to go deeper in regions of specific interest but should not drive design.

Field

Factors affecting choice of Fields

• Factors affecting quality of data– Cirrus Confusion Noise– Zodiacal photon noise

• Factors affecting the ease of conducting the survey– FIRST visibility– Existing survey data– Easy of ground based follow-up

Cirrus Confusion

5.1

1

5.25.2

1MJysr1mm1001mJy

BD

μ

Equating

From Gautier et al. (1992, AJ 103, 1313) and

Helou & Beichman (1990, Proc. 29th Liege Int. Astro. Colloq. ESA SP-314 ).

sourcecirrus 3.420

Normalising to B100 using cirrus spectrum (Rowan-Robinson et al 1992,

MNRAS, 258, 787 )

Factor of ~5 issafety margin ensuring 2x better than Marano at 175m

Example: ISO 175m observations of Marano

4.3source=107 mJy

B100 = 0.88 MJy/sr

10cirrus = 116 mJy

1/(Effective beam radius = 30.6’’)

Lagache and Puget 2000 A&A (astro-ph/9910255))

Arguably the limit at which you believe distinction between confused sources and cirrus

Sources extracted to 100 mJy

Cirrus Confusion limits

m 70 120 175 250 350 500D/m 3.5 3.5 3.5 3.5 3.5 3.5

B100/MJy/sr 2.36 1.85 2.34 3.18 4.23 5.81

Good Visibility

• Ease of scheduling FIRST survey observations– Ease of FIRST follow-up observations– Minimum impact on other FIRST science– Flexibility for orientation of maps

• Good visibility for other satellite observations

Visibility Constraints

• Solar elongation – >60 and <120

• Consider this over a year

• ||>45 gives visibility >50%

• Lower visibility is possible but a number of fields should then be distributed in so that some fields are always visible

Visibility & Cirrus ConstraintsIRAS 100m Cirrus map from Schlegel et al. 1998

B100 Contours at 1 and 2 MJy/sr

||=45 contours

Practical need for other survey data

• Degeneracy between T and z means FIRST data on it own is limited

• Large error circle and large dispersion between FIR and other bands mean identification difficult

• more bands decreases number of IDs

SdS

dndS

dS

dnSSPN IDs ~)|( FIR

)|()|,( FIRFIR2 SSPSSSP

Criteria for other surveys

• Area > 10 sq. deg. – smaller fields can be tackled individually on

case by case basis

• Area < 10,000 sq. deg– larger surveys do not constrain the fields

• Flux limits– minimum to detect at least half the objects– deeper surveys would of course be much better

The First FIRST source

• SED– FIR starburst from Efstathiou & Rowan-Robinson– Optical SED from Bruzual & Charlot 1996– X-ray M82 from Tsuru et al. 199– X-ray Sy1/2 from Barcons et al. 1995

– Radio, S=0.8

• Normalisations– LB|L60 from Saunders et al. 1990 at L60=1011

– X-ray using S15 of M82

– Radio, S1.4GHz/S90~100

• Source– z = 1(approximate median redshift of Rowan-Robinson 2000)

– S250 = 18.6 mJy (confusion limit from Rowan-Robinson 2000)

The First FIRST source

z=1

S250 = 18.6mJy

Existing/Potential Surveys

• SIRTF SWIRE Legacy Programme Lonsdale et al.– ~ 70 sq. deg at all SIRTF photom. bands

– Constraints more severe than for FIRST

– should be able to detect first FIRST source in IRAC bands IRAC MIPS

3.6 4.5 5.8 8.0 24 70 160

5 7.3 Jy

9.7 Jy

27.5 Jy

32.5 Jy

0.45 mJy

2.75 mJy

17.5 mJy

Existing/Potential Surveys

• XMM-LSS– 5x10-15 erg cm2 s-1

– 64 square degrees (low-– should detect first FIRST source if a Seyfert 1/2 not if

star-bust

• GALEX– 200 square degrees UAB =26 - fields?

• Radio– needs to be ~100Jy or better over 100 sq deg.?

Existing/Potential Surveys

• ESO-VIRMOS– Fields scattered making total of 16deg^2

• NOAO– SIRTF Legacy follow-up

• Follow-up of XMM-LSS fields 64 sq deg. CFHT & VLT• UKIDSS

– 100 sq deg K=21 (J, H, to similar depth)

• ESO VST– commitment to SIRTF Legacy & XMM-LSS

• VISTA– 250 sq deg. g’=28, r’=26.7, i’=26.2, z’=24.5, J=23.5, H=22.5, K=22

Conclusions

• Major survey project with FIRST is likely to be a confusion limited 100 sq deg.

• FIRST specific constraints are not severe• Complementary surveys will be very important

to science of FIRST• Need to actively ensure that surveys planned

now are suitable for FIRST• SIRTF SWIRE Legacy fields are likely to be

most appropriate